Here, then, is a revised version of Marshall's rules: (1) Figure out what you think about an issue, working back and forth among verbal intuition, evidence, and as much math as you need. (2) Stay with it till you are done. (3) Publish the intuition, the math, and the evidence - all three - in an economics journal. (4) But also try to find a way of expressing the idea without the formal apparatus. (5) If you can, publish that where it can do the world some good.
"Two Cheers for Formalism", The Economic Journal, Vol. 108, No. 451 (Nov., 1998)